Table 3.
Estimated effects of change in medication and number of paired glucose tests on A1c level and instantaneous linear change in A1c at baseline and at 3 and 6 months.
| Predictor | Estimated effect on A1c levela | Estimated effect on the instantaneous linear change rate in A1c | |||||
|
|
MLEb | t-ratio (df)c | P | MLEb | t-ratio (df)c | P | |
| Medication change |
|
|
|
|
|
|
|
|
|
3 months | –1.05 | –3.42 (157) | <.001 | 0.35 | 1.81 (157) | .07 |
|
|
6 months | –0.71 | –3.37 (157) | <.001 | 0.35 | 1.81 (157) | .07 |
| Number of paired tests |
|
|
|
|
|
|
|
|
|
3 months | –0.007 | –0.99 (156) | .33 | –0.008 | –1.82 (156) | .07 |
|
|
6 months | –0.015 | –1.86 (156) | .06 | –0.008 | –1.82 (156) | .07 |
a Models were estimated in a hierarchical fashion.
b MLE: maximum-likelihood estimate.
c t-ratios are ratios of the estimates to their respective standard errors.